[clang] [llvm] [aarch64] Add more MSVC intrinsics (PR #193221)

[Adam Glass via llvm-commits]

AdamGlass wrote:

"I expect almost everyone just uses standard C/C++ atomics, or the Win32 APIs; there isn't really much reason to drop down into raw intrinsics."

It happens.  Because of the n different ways of doing atomic operations on Aarch64, one of them may have some advantage (performance typically) for a specific case that doesn't generalize to generic atomics.  These intrinsics are a way to force use of a specific variant in a specific code path.

https://github.com/llvm/llvm-project/pull/193221